International Journal of Molecular Sciences

Article The Antimalarial Mefloquine Shows Activity against Mycobacterium abscessus, Inhibiting Mycolic Acid Metabolism

Giulia Degiacomi 1,†, Laurent Roberto Chiarelli 1,† , Deborah Recchia 1 , Elena Petricci 2 , Beatrice Gianibbi 2 , Ersilia Vita Fiscarelli 3 , Lanfranco Fattorini 4 , Fabrizio Manetti 2 and Maria Rosalia Pasca 1,*

1 Department of Biology and Biotechnology “Lazzaro Spallanzani”, University of Pavia, 27100 Pavia, Italy; [email protected] (G.D.); [email protected] (L.R.C.); [email protected] (D.R.) 2 Department of Biotechnology, Chemistry and Pharmacy, University of Siena, via Aldo Moro 2, 53100 Siena, Italy; [email protected] (E.P.); [email protected] (B.G.); [email protected] (F.M.) 3 Cystic Fibrosis Diagnostics, Microbiology and Immunology Diagnostics, Bambino Gesù Children’s Hospital IRCCS, 00165 Rome, Italy; evita.fi[email protected] 4 Dipartimento di Malattie Infettive, Istituto Superiore di Sanità, 00161 Rome, Italy; [email protected] * Correspondence: [email protected]; Tel.: +39-0382-985576 † These authors contributed equally to this work.

Abstract: Some nontuberculous mycobacteria (NTM) are considered opportunistic pathogens. Nev-

 ertheless, NTM infections are increasing worldwide, becoming a major public health threat. Further-  more, there is no current specific drugs to treat these infections, and the recommended regimens

Citation: Degiacomi, G.; Chiarelli, generally lack efficacy, emphasizing the need for novel antibacterial compounds. In this paper, we L.R.; Recchia, D.; Petricci, E.; Gianibbi, focused on the essential mycolic acids transporter MmpL3, which is a well-characterized target of B.; Fiscarelli, E.V.; Fattorini, L.; several antimycobacterial agents, to identify new compounds active against Mycobacterium abscessus Manetti, F.; Pasca, M.R. The (Mab). From the crystal structure of MmpL3 in complex with known inhibitors, through an in silico Antimalarial Mefloquine Shows approach, we developed a pharmacophore that was used as a three-dimensional filter to identify Activity against Mycobacterium new putative MmpL3 ligands within databases of known drugs. Among the prioritized compounds, abscessus, Inhibiting Mycolic Acid mefloquine showed appreciable activity against Mab (MIC = 16 µg/mL). The compound was con- Metabolism. Int. J. Mol. Sci. 2021, 22, firmed to interfere with mycolic acids biosynthesis, and proved to also be active against other NTMs, 8533. https://doi.org/10.3390/ijms including drug-resistant clinical isolates. Importantly, mefloquine is a well-known antimalarial agent, 22168533 opening the possibility of repurposing an already approved drug, which is a useful strategy to reduce the time and cost of disclosing novel drug candidates. Academic Editor: Patricia Agudelo-Romero Keywords: Mycobacterium abscessus; MmpL3; pharmacophore model; mefloquine; drug repurposing

Received: 13 July 2021 Accepted: 6 August 2021 Published: 8 August 2021 1. Introduction Publisher’s Note: MDPI stays neutral Nontuberculous mycobacteria (NTM) pulmonary infections are an emergent global with regard to jurisdictional claims in health threat, particularly to individuals that have a predisposing medical condition such published maps and institutional affil- as cystic fibrosis (CF), chronic obstructive pulmonary disease (COPD), and bronchiectasis. iations. Over the last two decades, the incidence of NTM infections among CF individuals has raised from 5% to 20%, increasing morbidity and mortality associated with these pathogens [1,2], whereas COPD patients treated with inhaled corticosteroid therapy are associated with a 29-fold increased risk of NTM infections [3]. NTM comprise more than 190 species and Copyright: © 2021 by the authors. subspecies with only a few causing serious and often opportunistic infections in humans. Licensee MDPI, Basel, Switzerland. Amongst them, the most commonly identified species in CF individuals are the slow This article is an open access article growing Mycobacterium avium complex (MAC) and the rapidly growing Mycobacterium distributed under the terms and abscessus complex (MABSC), accounting for 95% of CF cases [4]. MAC mainly consists of conditions of the Creative Commons M. avium subsp. avium (Mav) and M. avium subsp. intracellulare, while MABSC includes the Attribution (CC BY) license (https:// following three subspecies: M. abscessus subsp. abscessus (Mab), M. abscessus subsp. bolletii, creativecommons.org/licenses/by/ and M. abscessus subsp. massiliense. 4.0/).

Int. J. Mol. Sci. 2021, 22, 8533. https://doi.org/10.3390/ijms22168533 https://www.mdpi.com/journal/ijms Int. J. Mol. Sci. 2021, 22, 8533 2 of 16

A standard of care treatment for NTM is available for only a few pathogens, including MAC and Mab. Generally, the therapeutic options consist of multiple antimicrobial agents that are often associated with clinically significant adverse reactions and must be adminis- tered for prolonged periods [5]. A multidrug regimen based on a ( or ) is the recommended treatment for MAC, which also includes and ethambutol. Meanwhile, the management of pulmonary infections originated by Mab typically consists of an oral macrolide in conjunction with intravenous or inhaled , and one or more of the following drugs: intravenous cefoxitin, imipenem, or , in addition to oral (, clofazimine, moxifloxacin, and ) [6]. The American Thoracic Society and the Infectious Diseases Society of America stated that the optimal drug regimens and duration of anti-Mab treatment are not known. Moreover, treatment outcomes are poor, and reinfection with another strain or species is common [5]. The main causes of treatment failure are the intrinsic and acquired resistance of NTM in general, and of Mab in particular, which is one of the most drug-resistant mycobacteria [7]. Furthermore, the insufficient effectiveness of drugs toward Mab is associated with a lack of bactericidal activity of most of the active compounds [3,8]. In this context, there is an urgent need of more active drugs to curb NTM infections. Nevertheless, de novo drug discovery is a highly costly and time-consuming pro- cess [3]. Therefore, the study of approved drugs for new therapeutical applications provides a rapid strategy to clinical implementation. The drug pipeline for NTM is nearly empty with lead compounds derived either from the repurposing of existing drugs or from the cross-testing of compounds active in tuberculosis [9], but also includes a few new chemical entities. Interestingly, some of the novel drug candidates target MmpL3 (mycobacterial membrane protein Large 3) [10–14]. MmpL3 is an essential inner membrane flippase present in all mycobacteria, which is involved in the last step of mycolic acid biosynthesis. In particular, MmpL3 transports by flipping trehalose monomycolates across the inner membrane into the periplasm, where they are required to form the outer membrane mycolic acids [15,16]. The MmpL3 therapeu- tic potential has already been unveiled in several studies as a vulnerable and promiscuous target in Mycobacterium tuberculosis [4]. In this study, we combined an in silico approach (developing a pharmacophore, from the crystal structure of MmpL3 in complex with the inhibitors AU1235, SQ109, ICA38, and rimonabant [17]) with the repurposing strategy. Our pharmacophore was exploited for the screening of the DrugBank and the National Cancer Institute database, in order to identify novel scaffolds targeting MmpL3 belonging to molecules currently in use, for a future implementation as anti-Mab compounds. Amongst them, the disclosure of mefloquine, an antimalarial, was of particular interest.

2. Results 2.1. Molecular Modeling: Construction of the Pharmacophore The X-ray crystallographic three-dimensional structures of the complexes between MmpL3 and AU1235, SQ109, ICA38, and rimonabant [17] (Figure1) were used to generate four different pharmacophoric models by means of the software Phase [18]. Int. J. Mol. Sci. 2021, 22, 8533 3 of 16 Int. J. Mol. Sci. 2021, 22, x FOR PEER REVIEW 3 of 16

FigureFigure 1. StructureStructure of of the the MmpL3 MmpL3 inhibitors inhibitors used used to tobuild build the the pharmacophoric pharmacophoric model. model. The Thetri- tri- fluorophenylfluorophenyl moiety of AU1235, AU1235, the the difluorophenyl difluorophenyl moiety moiety of of ICA38, ICA38, the the terminal terminal alkenyl alkenyl edge edge of ofSQ109, SQ109, and and the the5-chlorophenyl 5-chlorophenyl moiety moiety of rimonabant of rimonabant are represented are represented in red. in The red. adamanty The adamantyll moi- ety of both AU1235 and SQ109, the spiro-undecane moiety of ICA38, and the piperidine ring of moiety of both AU1235 and SQ109, the spiro-undecane moiety of ICA38, and the piperidine ring of rimonabant are represented in green. rimonabant are represented in green.

Receptor–ligandReceptor–ligand interactions interactions were were coded coded into into automatically automatically generated generated pharmacophores pharmaco- thatphores also that comprised also comprised receptor-based receptor excluded-based excluded volumes. volumes. The four The pharmacophores four pharmacophores were thenwere mergedthen merged into a into common a common feature feature pharmacophoric pharmacophoric hypothesis hypothesis constituted constituted by the by sole the chemicalsole chemical features featur sharedes shared by the by parent the parent models. models. The resulting The resulting five-feature five-feature pharmacophore pharmaco- showsphore threeshows hydrophobic three hydrophobic regions regions (represented (represented by green by spheres green inspheres Figure 2in), Figure one aromatic 2), one ringaromatic (the orangering (the circle), orange and circle), a hydrogen and a hydrogen bond donor bond (the donor cyan (the sphere), cyan sphere), in addition in addi- to excludedtion to excluded volumes volumes (the transparent (the transparent cyan spheres cyan spheres of variable of variable size) that size) represent that represent regions ofregions space of occupied space occupied by the by amino the amino acid side acid chains side chains and areand thus are thus forbidden forbidden to portions to portions of theof the ligand ligand structures. structures. A A comparison comparison between between the the pharmacophore pharmacophore and and thethe MmpL3MmpL3 bind-bind- inging site s showedhowed that that the the vicinal vicinal hydrophobic hydrophobic portions portions and and the the aromatic aromatic ring ring feature feature of the of themodel model are accommodated are accommodated within within a pocket a pocket (delimited (delimited by the byside the chains side of chains Leu642, of Leu642, Val638, Val638,Leu708, Leu708, Ile249, Ile249,and Ile253) and Ile253),, where where the tri the-fluorophenyl tri-fluorophenyl moiety moiety of AU1235 of AU1235 (the (themost most im- importantportant molecular molecular portions portions of the of the four four MmpL3 MmpL3 inhibitors inhibitors are arehighlighted highlighted in Figure in Figure 1), 1the), thedi-fluorophenyl di-fluorophenyl moiety moiety of of ICA38, ICA38, the the 5- 5-chlorophenylchlorophenyl moiety moiety of of rimonabant, and the the alkenylalkenyl edge of SQ109 SQ109 are are found found in in the the corresponding corresponding crystallographic crystallographic complexes. complexes. On Onthe theother other hand, hand, the hydrogen the hydrogen bond bond acceptor acceptor feature feature of the of model the model takes takes into account into account the pres- the presenceence of amide of amide or amine or amine NH NHgroups groups of the of theinhibitors inhibitors that that give give hydrogen hydrogen bonds bonds with with the theside side chain chain of Asp645. of Asp645. The remaining The remaining hydrophobic hydrophobic feature feature of the ofmodel the modelrepresents represents the ad- theamantyl adamantyl moiety moiety of both of both AU12345 AU12345 and andSQ109, SQ109, as well as well as the as the spiro spiro-undecane-undecane moiety moiety of ofICA38 ICA38 or orthe the piperidine piperidine ring ring of of rimonabant, rimonabant, and and it it is is located located in a pocketpocket delimiteddelimited byby Tyr257,Tyr257, Phe649,Phe649, andand Tyr646.Tyr646.

Int. J. Mol. Sci. 2021, 22, 8533 4 of 16 Int. J. Mol. Sci. 2021, 22, x FOR PEER REVIEW 4 of 16

FigureFigure 2. GraphicalGraphical representation representation of of the the final final five five-feature-feature pharmacophoric pharmacophoric model model for MmpL3 for MmpL3 inhibitorsinhibitors embeddedembedded withinwithin thethe bindingbinding sitesite thatthat contains contains AU1235. AU1235. AminoAmino acidsacids are are represented represented by thinby thin lines lines (atom (atom type type notation), notation), while while the the inhibitor inhibitor is renderedis rendered by by thick thick lines lines (atom (atom type type notation). nota- Greention). Green spheres spheres represent repr theesent hydrophobic the hydrophobic features features of the of pharmacophore, the pharmacophore, the orange the orange circle circle is the is the aromatic ring, while the cyan sphere corresponds to the hydrogen bond acceptor feature. Light aromatic ring, while the cyan sphere corresponds to the hydrogen bond acceptor feature. Light cyan spheres of variable size represent excluded volumes. Hydrogen bonds between the amidic cyan spheres of variable size represent excluded volumes. Hydrogen bonds between the amidic NH NH groups of the inhibitor and Asp645 side chain are coded by dashed yellow lines. groups of the inhibitor and Asp645 side chain are coded by dashed yellow lines. The final pharmacophore was used as a three-dimensional filter to identify 36 puta- The final pharmacophore was used as a three-dimensional filter to identify 36 putative tive MmpL3 ligands within databases of known drugs (such as the DrugBank database) MmpL3 ligands within databases of known drugs (such as the DrugBank database) [19] and commercially[19] and commercially available available compounds compounds (such as the(such National as the National Cancer Institute Cancer database)Institute data- [20] (Tablebase) [20] S1). (Table S1).

2.2.2.2. CharacterizationCharacterization ofof MmpL3MmpL3 LigandsLigands ActiveActive againstagainst MabMab GrowthGrowth. AsAs aa result,result, amongamong thethe 3636 prioritizedprioritized compounds,compounds, NSC157387NSC157387 (mefloquine,(mefloquine, 11),), thethe tricyclictricyclic urea urea NSC120330, NSC120330, and and the the benzimidazole benzimidazole NSC135792 NSC135792 are active are against active againstMab growth Mab withgrowth an with MIC an = MIC 16 µg/mL = 16 μg (Table/mL (Table S1 and S1 and Table Table1). As 1). anAs examplean example of of a matcha match between between thethe pharmacophorepharmacophore andand thethe chemicalchemical portionsportions ofof thethe ligands,ligands, thethe trifluoromethylphenyltrifluoromethylphenyl moietymoiety (the(the mostmost importantimportant molecularmolecular portionsportions ofof mefloquinemefloquine structurestructure areare highlightedhighlighted inin FigureFigure S1)S1) ofof mefloquinemefloquine matchedmatched thethe twotwo vicinalvicinal hydrophobichydrophobic regionsregions (Figure(Figure3 3),), while while thethe pyridinepyridine ring ring corresponded corresponded to to the the aromatic aromatic ring ring feature. feature. The The basic basic nitrogen nitroge atomn atom in its in protonatedits protonated form form was was the hydrogenthe hydrogen bond bond donor, donor and, theand remaining the remaining hydrophobic hydrophobic feature fea- is mappedture is mapped by the C2-C3by the alkylC2-C3 portion alkyl portion of the piperidine of the piperidine nucleus. nucleus.

Int. J. Mol. Sci. 2021, 22, x FOR PEER REVIEW 5 of 16 Int. J. Mol. Sci. 2021, 22, x FOR PEER REVIEW 5 of 16 Int. J. Mol. Sci. 2021, 22, 8533 5 of 16

Table 1. Structure and anti-Mab activity of 3 active MmpL3 ligands and their analogs. TableTable 1. Structure1. Structure and and anti anti-Mab-Mab activity activity of 3of active 3 active MmpL3 MmpL3 ligands ligands and and their their analogs. analogs. Table 1. Structure and anti-Mab activity of 3 active MmpL3 ligands and their analogs.

MIC Against MIC AgainstMIC Against 1 2 3 MIC Against Compound R R11 R22 R3 M.3 abscessus ATCC 19977 CompoundCompound R R R R1 RR R2R R3 RM.M. abscessus abscessusM. abscessus ATCC ATCCATCC 19977 19977 19977 (μg/mL)(µg/mL) a a (μg(μg/mL)/mL) a NSC120330 16 NSC120330NSC120330 16 16 16 NSC135792NSC135792 16 16 NSC135792NSC135792 16 16

NSC157387,NSC157387, mefloquine, mefloquine, 11 CFCF3 3 HH CF3 CF3 16 16 NSC157387, mefloquine, 1 CF3 H CF3 16

NSC305803,NSC305803, 22 CFCF3 HH CF3 CF >64 >64 NSC305803, 2 CF33 H CF3 3 >64

NSC305814,NSC305814, 33 CFCF3 --CHCH2NH(CHNH(CH 2))2CHCH3 HH CF3 CF >64 >64 NSC305814, 3 CF33 -CH2 2NH(CH2 2 2)2CH3 3 H CF3 3 >64 NSC305816, 4 CF3 -CH2N(CH-2CH2CH3)2 H CF3 >64 NSC305816, 4 4 CF3 -CH2N(CH2CH2CH3)2 HH CF CF3 >64>64 3 CH N(CH CH CH ) 3 NSC305815, 5 CF3 -CH2 2NH2 C(CH2 3)33 2 H CF3 >64 NSC305815, 5 CF3 -CH2NH C(CH3)3 H CF3 >64 NSC305815, 5 CF3 -CH2NH C(CH3)3 H CF3 >64 NSC305817, 6 CF3 -(CH2)2NH(CH2)2CH3 H CF3 >64 NSC305817, 6 CF3 -(CH2)2NH(CH- 2)2CH3 H CF3 >64 NSC305817, 6 CF3 2 2 2 3 3 2 H3 CF >64 NSC305793, 7 CF3 3 -(CH2)2N((CH2)3CH3)2 H CF3 3 32 NSC305793, 7 CF3 (CH-(CH2)2NH(CH2)2N((CH2)22CH)3CH3 3)2 H CF3 32 - NSC305793, 7 CF3 H CF3 32 NSC158680, 8 CF3 (CH2)2N((CH2)3CH3)2 H F >64 NSC158680, 8 CF3 H F >64

NSC158680, 8 CF3 H F >64 NSC4377, 9 p-Cl-Phenyl H Cl 16 NSC4377,NSC4377, 9 9 p-Clp--ClPhenyl-Phenyl H HCl Cl 16 16

NSC4377, 9 p-Cl-Phenyl 2 2 3 3 2 H Cl 16 NSC369051, 10 p-Cl-Phenyl -CH2N((CH2)3CH3)2 Cl Cl >64 NSC369051, 10 p-Cl-Phenyl -CH2N((CH2)3CH3)2 Cl Cl >64

NSC305758,NSC369051, 1110 Phenylp-Cl-Phenyl -CH2N((CH2)3CH3)2 ClCl Cl Cl32 >64 NSC305758,NSC305758, 11 11 PhenylPhenyl Cl ClCl Cl 32 32

NSC305758, 11 Phenyl Cl Cl 32 NSC13446, 12 cyclohexyl H H >64 NSC13446,NSC13446, 12 12 cyclohexylcyclohexyl H H H H >64>64

NSC13446, 12 cyclohexyl H H >64 NSC13480, 13 Phenyl 16 NSC13480,NSC13480, 13 13 PhenylPhenyl 16 16

a NSC13480, 13 a DataPhenyl representative of three independent experiments. 16 Dataa Data representative representative of threeof three independent independent experiments. experiments.

a Data representative of three independent experiments.

Int. J. Mol. Sci. 2021 22 Int. J. Mol. Sci. 2021, 22, , x ,FOR 8533 PEER REVIEW 6 of6 of16 16

Figure 3. Graphical representation of mefloquine superimposed to the pharmacophoric model. The Figure 3. Graphical representation of mefloquine superimposed to the pharmacophoric model. hydrophobic features (green spheres) are matched by the trifluoromethylphenyl moiety (the most The hydrophobic features (green spheres) are matched by the trifluoromethylphenyl moiety (the mostimportant important molecular molecular portions portions of the of mefloquinethe mefloquine structure structure are highlightedare highlighted in Figure in Figure S1) and S1) theand alkyl theportion alkyl portion of the piperidine of the piperidine ring, the ring, aromatic the aromatic ring feature ring (orange feature circle) (orange is mapped circle) is by mapped the pyridine by the core, pyridineand the core, hydrogen and the bond hydrogen donor bond group donor (cyan group sphere) (cyan is represented sphere) is represented by the basic by piperidine the basic nitrogenpi- peridineatom. Excludednitrogen atom. volumes Excluded (light cyanvolumes spheres (light of cyan variable spheres size) of are variable also displayed. size) are also displayed.

ToTo further further support support the the hypothesis hypothesis that that such such compounds compounds were were able able to tobind bind MmpL3, MmpL3, theythey were were submitted submitted to tomolecular molecular docking docking simulations simulations and and minimization minimization of of the the resulting resulting complexes.complexes. An An analysis analysis of ofthe the best best-docked-docked pose pose of ofmefloquine mefloquine within within the the MmpL3 MmpL3 binding binding sitesite showed showed extended extended hydrop hydrophobichobic contacts contacts between between the the trifluoromethyl trifluoromethyl group group at atC8 C8 and and thethe side side chains chains of ofVal638 Val638 and and Ile297 Ile297 (Figure (Figure 4),4), as as well well as as an an additional additional interaction interaction between between thethe condensed condensed phenyl phenyl ring ring and and Leu642. Leu642. Another Another hydrophobic hydrophobic pocket pocket comprising comprising Ile249, Ile249, Ile253,Ile253, and and Leu686 Leu686 accommodated accommodated the the 2- 2-trifluromethylpyridyltrifluromethylpyridyl moiety. moiety. Moreover, Moreover, Ile253 Ile253 alsoalso interacted interacted with with the the alkyl alkyl portion portion of ofthe the piperidine piperidine ring, ring, together together with with Tyr257 Tyr257 and and Tyr646.Tyr646. The The most most important important anchor anchor point point of mefloquine of mefloquine was wasrepresented represented by the by basic the basicni- trogennitrogen atom atom of th ofe thepiperidine piperidine nucleus nucleus that that was was able able to tomake make a hydrogen a hydrogen bond bond with with the the sideside chain chain of of Ser293 Ser293 and and an an additional additional charge charge-reinforced-reinforced hydrogen bondbond withwith thethe carboxylic carbox- ylicedge edge of of Asp645, Asp645, found found to to play play aa pivotalpivotal rolerole for the interaction interaction of of MmpL3 MmpL3 with with all all the the coco-crystallized-crystallized ligands. ligands. A Abi- bi-dimensionaldimensional diagram diagram of ofthe the interactions interactions between between mefloquine mefloquine andand amino amino acid acid residues residues mostly mostly contributing toto the the stabilization stabilization of of the the MmpL3-mefloquine MmpL3-meflo- complex, and distances between molecular portions of mefloquine and amino acid groups quine complex, and distances between molecular portions of mefloquine and amino acid are depicted in Figure S2. groups are depicted in Figure S2.

Int.Int. J.J. Mol.Mol. Sci.Sci. 2021,, 22,, 8533x FOR PEER REVIEW 77 of 1616

Figure 4. Graphical representation of the interaction pattern of the best-dockedbest-docked pose of mefloquinemefloquine within the MmpL3 binding site. Extended Extended hydrophobic hydrophobic contacts contacts involve involve both both the the trifluoromethyl trifluoromethyl substituents andand thethe quinolinequinoline core. core. The The piperidine piperidine basic basic nitrogen nitrogen atom atom makes makes a hydrogena hydrogen bond bond with thewith Ser293 the Ser293 side chains,side chains, and aand charge-reinforced a charge-reinforced hydrogen hydrogen bond bond with with the carboxylthe carboxyl edge edge of Asp645. of Asp645. Amino acids are represented by sticks (atom type notation), while mefloquine is repre- Amino acids are represented by sticks (atom type notation), while mefloquine is represented by a ball sented by a ball and stick (atom type notation). Yellow dashed lines codify for hydrogen bonds, and stick (atom type notation). Yellow dashed lines codify for hydrogen bonds, while the magenta while the magenta dashed line represents a charge-reinforced hydrogen bond. dashed line represents a charge-reinforced hydrogen bond.

2.3. SAR ConsiderationsConsiderations of the 3 Anti-MabAnti-Mab MmpL3 LigandsLigands A SAR analysis on a small library of anan additionaladditional 1212 mefloquinemefloquine derivatives (2–132–13) provided by thethe NationalNational CancerCancer InstituteInstitute suggestedsuggested that the basicbasic piperidinepiperidine ringring (the(the black portion in Figure S1) atat thethe appropriateappropriate distancedistance fromfrom thethe quinolinequinoline nucleusnucleus (the(the green portion in Figure S1)S1) waswas mandatorymandatory forfor itsits activityactivity againstagainst MabMab wild-typewild-type strainstrain (Table(Table1 1).). In fact, the opening of the piperidine ring (33––77)) oror lengtheninglengthening ofof thethe linkerlinker betweenbetween the two cycles cycles ( (22)) resulted resulted in in the the loss loss of of antimycobacterial antimycobacterial activity. activity. On On the theother other hand, hand, the thereplacement replacement of the of C8 the trifluoromethyl C8 trifluoromethyl moiety moiety with a with chloride a chloride maintained maintained activity activity only if onlythe piperidine if the piperidine ring was ring kept was intact. kept In intact. fact, 9 Inand fact, 11 showed9 and 11 a showedMIC of 16 a MICand 32 of μg 16/mL, and 32respectively,µg/mL, respectively, while 10 was while inactive.10 was Moreover, inactive. Moreover, further reduction further reductionss in the size in theand size hydro- and hydrophobicphobic character character of the ofC8 the substituent C8 substituent to F and to F H and led H to led inactive to inactive compounds compounds (8 and (8 and12). 12Finally,). Finally, the thecondensed condensed derivative derivative 13 showed13 showed an aninteresting interesting antimycobacterial antimycobacterial activity activity of of16 16 μgµ/mL.g/mL. The The binding binding pose pose and and the the interaction interaction pattern pattern hypot hypothesizedhesized for mefloquine mefloquine agreed with the major SAR found found for for its its congeners. congeners. In In fact, fact, mefloquine mefloquine derivatives derivatives with with a alinear linear or or branched branched amine amine side side chain chain (3– (73)– underwent7) underwent conformational conformational rearrangements rearrangements that that led to the lack of the pivotal hydrogen bonds with Asp645 and Ser293. Moreover, a

Int. J. Mol. Sci. 2021, 22, 8533 8 of 16

significant reduction in the size and lipophilicity of the C8 substituent to F and H nullified the hydrophobic interactions found for the CF3 group of mefloquine and affected negatively the antimycobacterial activity. A small library of an additional 8 compounds provided by Enamine Ltd. (Kyiv, Ukraine) was also built around the tricyclic urea NSC120330; however, none of the assayed compounds were active (Table S2). This SAR trend could be reasonably justified by the structural properties of such compounds. In fact, NSC120330 showed the presence of an alkylating edge that lacked in the congeneric compounds. Consequently, we can assume that the antimycobacterial activity of NSC120330 was not target-specific and was derived from its alkylating properties. On this basis, NSC120330 and its analogs were not further studied.

2.4. Analysis of Lipid Composition of Mab Cells To confirm that mefloquine and NSC135792 actually target MmpL3, we analyzed the global lipid profile and the mycolic acids content of Mab cells, upon treatment with both compounds. To this purpose, Mab cells were grown in the presence of 25× MIC of the compounds, harvested after 48 h of incubation, and the lipids were extracted and analyzed as reported in Material and Methods. As a positive control, cells were grown in the presence of 10X MIC of the well-known MmpL3 inhibitor BM212 [21]. As shown in Figure5, the global lipid profiles of Mab cells treated with both mefloquine and NSC135792 were very similar to those of the cells treated with BM212 (positive control). Mab cultures grown in the presence of the higher solvent concentration used (4% DMSO for mefloquine and NSC135792, 4% ethanol for BM212) as a negative control did not show significant differences with untreated cultures. It is known that in M. tuberculosis, BM212 acts to block the TDM transport by inhibiting MmpL3 [22]. Indeed, as it was nicely demonstrated by Xu et al., 2017 [21], its mechanism of action is not due to indirect effects on the proton motive force, but BM212 binds MmpL3, directly inhibiting its activity. In our cases, cells treated with NSC135792 and mefloquine, as well as with BM212, showed significant changes in the composition of the mycobacterial cell envelope in comparison with the untreated control, with at least two main spots altered, as indicated by the arrows in Figure5A. It was demonstrated that the effects of the tested compounds were concentration- dependent, being negligible at 1×-MIC concentration and more visible at 10×- and 25×- MIC concentrations (Figure S3). A comparable effect on lipid profile was found with BM212, suggesting that NSC135792 and mefloquine likely act through the inhibition of MmpL3 activity, although the mechanism of action of the pyrrole derivative has been demonstrated in M. tuberculosis, but not in Mab yet. Furthermore, mycolic acids were also extracted from delipidated bacteria to observe the possible impact on arabinogalactan (AG) mycolylation. The presence of both com- pounds led to a very significant decrease in α-mycolates (Figure5B), as also visible in the positive control sample (lane 2). The decrease we observed could be due to the block of TMM transport that could indirectly perturb AG mycolylation in all the treated samples. It is conceivable that TMM transport could be actually blocked by the studied com- pounds, as well as BM212 in M. tuberculosis. On the other hand, we cannot exclude that the observed phenotype could be the result of a possible inhibition in different steps of mycolic acids biosynthesis, specifically of the latter condensation steps catalyzed by the FAS-II system. In fact, the compounds under investigation affected only one form of mycolates, the α-mycolates (Figure5C). Nevertheless, the two compounds (mefloquine and NSC135792) behave overall in a similar way with respect to the control BM212, suggesting that they possibly share the same mechanism of action; however, at this stage of the study, we can assume that an inhibition in different steps of mycolic acids biosynthesis occur, causing the observed phenotype. Int. J. Mol. Sci. 2021, 22, 8533 9 of 16 Int. J. Mol. Sci. 2021, 22, x FOR PEER REVIEW 9 of 16

FigureFigure 5. 5. EffectsEffects of of the the compounds compounds on on MabMab cellcell wall wall composition. composition. MabMab cellscells were were grown grown for for 48 48 h h in in thethe absence/presence absence/presence of of compounds, compounds, and and the the lipid lipid composition composition was was then then analyzed analyzed as asin Material in Material and Methods. (A) TLC analysis of the complete lipid profile extracted and separated in solvent and Methods. (A) TLC analysis of the complete lipid profile extracted and separated in solvent CHCl3/CH3OH/H2O (20:4:0.5). (B) total mycolic acid methyl esters (MAMEs) and fatty acid methyl CHCl /CH OH/H O (20:4:0.5). (B) total mycolic acid methyl esters (MAMEs) and fatty acid methyl esters 3(FAME)3 extracted2 from delipidated cell pellets were separated in hexane/ethyl acetate (95:5). estersand α (FAME)’-mycolic extracted acids correspond from delipidated to long- cellchain pellets (C77 were–79) and separated short-chain in hexane/ethyl (C62–64) mycolic acetate acids (95:5). inα Maband α[10]’-mycolic. Lane 1: acids no addition; correspond lane to 2: long-chain 10× MIC BM212; (C77–79) lane and 3: short-chain 25× MIC NSC135792; (C62–64) mycolic lane 4: acids 25× in MICMab mefloquine[10]. Lane 1:(4). no Figures addition; are lanerepresentative 2: 10× MIC of BM212;three independent lane 3: 25× experiments.MIC NSC135792; (C) Structures lane 4: 25 × ofMIC αand mefloquine α’-mycolic (4). acids Figures [23] are. representative of three independent experiments. (C) Structures of α and α’-mycolic acids [23]. It was demonstrated that the effects of the tested compounds were concentration- dependent,2.5. Evaluation being of Antimycobacterial negligible at 1x- ActivityMIC concentration of Selected Compounds and more againstvisible NTMat 10x- and 25x- Clinical Isolates MIC concentrations (Figure S3). A comparable effect on lipid profile was found with BM212,We suggesting assessed the that sensitivity NSC135792 to mefloquineand mefloquine and NSC135792likely act through against the a panelinhibition of both of MmpL3reference activity, strains andalthough drug-resistant the mechanism clinical of isolates action belonging of the pyrrole to MABSC derivative and MAC. has been The demonstratedmicro-broth dilution in M. tuberculosis method was, but used not toin determineMab yet. the MIC values. Furthermore,We tested, in mycolic addition acids to M. were abscessus also extractedATCC 19977, fromM. delipidated bolletii 1999-0888 bacteria and to observeM. mas- thesiliense possible2005-0524 impact (MABSC), on arabinogalactan alongside (AG)five Mab mycolylation.clinical isolates The presence resistant of to both at least com- a poundsquinolone led and to a a very , significant and decrease two Mab inmultidrug-resistant α-mycolates (Figure (MDR) 5B), as isolates, also visible one of in them the M. avium avium positiveisolated control from a CFsample patient. (lane Moreover, 2). The decrease wesubsp. observed Chestercould be ATCC15769 due to the referenceblock of strains (MAC) and four M. avium isolates resistant to at least linezolid were used (Table2). TMM transport that could indirectly perturb AG mycolylation in all the treated samples. Interestingly, mefloquine and NSC135792 were also active against other NTM strains It is conceivable that TMM transport could be actually blocked by the studied com- and NTM drug-resistant clinical isolates, also including Mab MDR strains. pounds, as well as BM212 in M. tuberculosis. On the other hand, we cannot exclude that the observed phenotype could be the result of a possible inhibition in different steps of mycolic acids biosynthesis, specifically of the latter condensation steps catalyzed by the

Int. J. Mol. Sci. 2021, 22, 8533 10 of 16

Table 2. Antimycobacterial activity of mefloquine and NSC135792 against NTM reference strains and drug-resistant clinical isolates.

MIC (µg/mL) Strains Genotype Provenience Mefloquine NSC135792 M. abscessus ATCC Wild-type 16 16 Reference strain 19977 M. bolletii 1999-0888 Wild-type 32 32 Hôpital Raymond-Poincaré M. massiliense (Paris, France) Wild-type 32 32 2005-0524 Resistant to AMK, CLR, DOX, Centre hospitalier M. abscessus MDR BDQ, CPFX, ERY, MEM, ECO, 32 32 universitaire vaudois clinical isolate 1 EMB, ETH, LAN, PRI, RIF, RIP, (Lausanne, Switzerland) SQ109, SZD, TAC Resistant to AMK, AMOX, CL, M. abscessus MDR Ospedale Pediatrico CFP, CFX, CTX, CPFX, DOX, clinical isolate 2 32 32 Bambino Gesù (Rome, IPM, LZD, MIN, MFX, TOB, (from CF patient) Italy) TMP-SMX M. abscessus Resistant to CLR, MFX, DOX, 32 32 ISS, Rome, Italy clinical isolate n. 6 LZD Resistant to CLR, AMK, MFX, M. abscessus DOX, 32 32 ISS, Rome, Italy clinical isolate n. 7 (Intermediate sensitivity to LZD) M. abscessus Resistant to MFX, DOX 32 32 ISS, Rome, Italy clinical isolate n. 8 M. abscessus Resistant to CLR, MOX, DOX 32 32 ISS, Rome, Italy clinical isolate n. 9 M. abscessus Resistant to MOX, DOX 32 32 ISS, Rome, Italy clinical isolate n.10 M. avium subsp. avium Wild-type 64 64 Reference strain Chester ATCC15769 Resistant to LZD M. avium (Intermediate sensitivity to 64 64 ISS, Rome, Italy clinical isolate n.1 MOX) Resistant to LZD M. avium (Intermediate sensitivity to 64 64 ISS, Rome, Italy clinical isolate n.2 MOX) Resistant to LZD M. avium (Intermediate sensitivity to 64 64 ISS, Rome, Italy clinical isolate n.3 MOX) M. avium Resistant to LZD and MOX 64 64 ISS, Rome, Italy clinical isolate n.4 Abbreviations: AMK amikacin, AMOX amoxicillin, BDQ bedaquiline, CL clavulanic acid, CFP cefepime, CFX cefoxitin, CLR clarithromycin, CTX ceftriaxone, CPFX ciprofloxacin, DOX , ECO econazole, EMB ethambutol, ERY , ETH ethionamide, IPM imipenem, LAN lansoprazole, LZD linezolid, MEM meropenem, MIN minocycline, MFX moxifloxacin, PRI pristinamycin, RIF rifampicin, RIP rifapentine, SQ109, SZD , TAC thiacetazone, TMP-SMX trimethoprim/sulfam, TOB .

2.6. Drug Combination with Mefloquine against Mab The potential inclusion of mefloquine in combinational Mab therapy was studied via Checkerboard Assay (CBA). NSC135792 was not further investigated, because the disclosure of a 35-compound library constituted 5,6-disubstituted-benzimidazole analogs with antimycobacterial activity as low as 0.125 µg/mL toward Mab (CIP104536) [24]. Be- daquiline, ciprofloxacin, and clarithromycin were tested. The main indicator of synergy, the Fractional Inhibitory Concentration Index (FICI), was >0.5–4 for each assayed antibi- otic, indicating no antagonism interaction with mefloquine and, thus, the suitability of mefloquine for a possible combinational therapy. Int. J. Mol. Sci. 2021, 22, 8533 11 of 16

3. Discussion Despite the increasing incidence in NTM infections, currently, there are no specific drugs against them. Moreover, the recommended regimens for the treatment of Mab lung infections generally lack efficacy [25], emphasizing the urgent need for new antibacterial compounds with novel mechanisms of action, also considering that very few molecules are in clinical development [8]. The cell wall biosynthesis pathways have represented a rich source of validated targets for antimycobacterial compounds. Mab does not represent an exception, as the majority of new compounds under investigation for its treatment are mycolic acid biosynthesis inhibitors. Among these targets, MmpL3 has emerged as the target of several different chemical entities showing antitubercular activity, being defined as a promiscuous target [26]. Interestingly, several of these inhibitors have been found active also against Mab or M. avium [4,14,27]. Moreover, novel Mab active compounds targeting MmpL3 have been identified through the phenotypic screen of compound libraries [10,28]. In this context, our study aimed to identify novel scaffolds, already commercially available, inhibiting MmpL3, that will be useful for the development of new anti-Mab compounds. To this purpose, we used an in silico approach, developing a pharmacophore, from the crystal structure of MmpL3 in complex with the antitubercular inhibitors AU1235, SQ109, ICA38, and rimonabant [17]. The screening of the DrugBank database and the National Cancer Institute database exploiting our pharmacophore afforded 36 prioritized compounds. Three of them (namely, mefloquine, the tricyclic urea NSC120330, and the benzimidazole NSC135792) showed an interesting activity against Mab growth (MIC = 16 µg/mL). In particular, we confirmed that mefloquine and NSC135792 interfere with mycolic acid biosynthesis, perturbing indirectly the AG mycolylation (Figure5B). It is plausible that the two compounds under investigation could share the same mechanism of action of the control and well-known MmpL3 inhibitor BM212, but it is not possible to exclude an inhibition downstream MmpL3, as a reduction of only α-mycolates was observed. Meanwhile, the indole-2-carboxamides and the piperidinol derivative PIPD1, two anti-Mab MmpL3 inhibitors, affect the formation of both α and α’ mycolates [8,9]. Interestingly, some analogs of NSC135792 have been recently published as very active compounds against Mab, inhibiting MmpL3 activity [23]. For this reason, this compound was not further investigated; nevertheless, this evidence supports the validity of our pharmacophore model. Another compound that emerged from our in silico analysis is the mefloquine, which is a well-known antimalarial compound [29]. This molecule is particularly interesting, opening the possibility of a drug repurposing strategy, an approach particularly considered for the antibacterials. Indeed, the emergence of antibiotic-resistance often overtakes the development of novel drugs, and in this context, the repurposing of already approved ones can be useful to reduce time and cost [8]. The characterization of the antimicrobial spectrum of mefloquine demonstrated that the compound is significantly active against other MABSC subspecies, including multidrug-resistant clinical isolates, and also displayed moderate activity against different MAC strains. Moreover, the compound did not show antagonistic effects in combination with other antimycobacterial drugs, as demonstrated by CBA, confirming that it could be used for the development of new drug combinations. In conclusion, with our study, we developed a pharmacophore model that allowed one to disclose novel scaffolds inhibiting the mycolic acids biosynthesis, endowed with significant anti-Mab activity. Furthermore, the identification of mefloquine among these compounds will allow further investigation for a repurposing strategy to disclose novel drug candidates. Int. J. Mol. Sci. 2021, 22, 8533 12 of 16

4. Materials and Methods 4.1. Computational Details Small molecules were sketched or imported in SMILES notation in the Maestro suite [30] and were prepared with the LigPrep routine, generating possible tautomers at pH 7 ± 1 with Epik [31] and using the OPLS3e force field. The X-ray crystallographic three-dimensional structure of MmpL3 in complex with AU1235 (entry 6ajh of the protein data bank, 2.8 Å resolution), SQ109 (entry 6ajg, 2.6 Å resolution), ICA38 (entry 6ajj, 2.8 Å resolution), and rimonabant (entry 6aji, 2.9 Å resolution) were imported from the protein data bank and submitted to the Protein Preparation Wizard by assigning bond orders and adding hydrogens to the appropriate positions. Next, starting from each receptor–ligand complex, a pharmacophore model was generated by means of the Develop Pharmacophore Model routine within the software Phase, adding a receptor-based excluded volume shell. The Hypothesis Alignment was then applied to finding feature matching between the four original pharmacophores, and only five pharmacophoric features shared by all the models were kept in the final pharmacophoric model. The latter was constituted by three hydrophobic regions, an aromatic ring feature, and a hydrogen bond donor group. The pharmacophore was used as a three-dimensional query to filter both the DrugBank (a com- prehensive, free-of-access, online database of about 14,000 entries, including small-molecule drugs, nutraceuticals, and experimental drugs) and the NCI/DTP Open Chemicals Repos- itory (a collection of more than 200,000 compounds provided at no cost upon request). Twenty-eight small molecules were prioritized as putative MmpL3 inhibitors and sub- mitted to biological evaluation. Three of them, namely, mefloquine (NSC157387), the arylurea NSC120330, and the 2-amino-5,6-dimethylbenzimidazole NSC135792, showed an antimycobacterial MIC of 16 µg/mL. In the attempt to find more active compounds and to enlarge SAR considerations, twelve additional mefloquine analogs were collected from the NCI Repository. On the contrary, the arylurea derivatives were abandoned as a consequence of the hypothesis that the activity of NSC120330 (a chloroethylurea) was based on its alkylating properties. In a similar way, the benzimidazole class was not further studied, because its members were already reported as inhibitors of M. abscessus [23]. To further support the hypothesis that mefloquine could represent a MmpL3 ligand, molecular docking simulations were also performed by means of the software Glide [32]. In particular, the Receptor Grid Generation module was applied to codify the physico- chemical properties of the MmpL3 binding site to be used during docking calculations. The grid box center (x, y, and z coordinates: 35.13, 3.69, −20.55) was embedded in a 15 × 15 × 15 Å inner box and a 35 × 35 × 35 Å outer box. All the receptor hydroxyl and thiol groups within the receptor grid were allowed to rotate. Next, a standard precision calculation with flexible ligand sampling was set with post-docking minimization of the resulting complexes.

4.2. Bacterial Strains, Culture Media, and Chemicals Mab ATCC 19977 and Mycobacterium avium subsp. avium Chester (ATCC15769) ref- erence strains were used alongside a series of clinical isolates belonging to Mab complex (MABSC) and M. avium complex (MAC) (Table2). These strains were grown at 37 ◦C in Middlebrook 7H9 broth (Difco, Becton Dickinson, NJ, USA), supplemented with 0.05% w/v Tween 80 or on Middlebrook 7H11 (Difco, Becton Dickinson, NJ, USA), both supplemented with 0.2% w/v glycerol and 10% v/v Middlebrook OADC enrichment (oleic acid, albumin, D-glucose, catalase; Becton Dickinson, NJ, USA). Mab MDR clinical strain n. 1 was isolated at the Centre hospitalier universitaire Vaudois (CHUV, Lausanne, Switzerland) and its antibiotic resistance profile was obtained by the resazurin-based microtiter assay (REMA). Mab MDR clinical isolate n. 2 was identified from a cystic fibrosis patient at the Bambino Gesù Children’s Hospital (Rome, Italy); the sensitivity to antibiotics was de- tected by the microbroth dilution method (RAPMYCO, Thermo Fisher Scientific, Waltham, MA, USA). Int. J. Mol. Sci. 2021, 22, 8533 13 of 16

The drug-resistant Mab and M. avium clinical strains were isolated at the Istituto Superiore di Sanità (Rome, Italy). The MIC values were determined by the Sensititre tests SLOMYCOI for M. avium e RAPMYCOI for M. abscessus. All the tested compounds were dissolved in DMSO (Sigma Aldrich, St. Louis, MO, USA). BM212 (Cayman Chemical, Ann Arbor, MI, USA) was dissolved in ethanol 70%.

4.3. MIC Determination Drug susceptibility of Mab was determined using the REMA method [33]. Log-phase cultures were diluted at concentrations of approximately 106 bacteria/mL. Then, 100 µL of the bacterial suspensions was added in a 96-well black plate (Fluoronunc, Thermo Fisher, Waltham, MA, USA) containing 100 µL of Middlebrook 7H9, without the addition of Tween 80, in the presence of serial compound dilution. Growth controls containing no compound and sterile controls without inoculum were also included. A volume of 10 µL of resazurin (0.025% w/v) was added to each well after 24 h, and bacterial viability was assessed after a further 18–24 h of incubation using a FluoroskanTM Microplate Fluorometer (Thermo Fisher Scientific, Waltham, MA, USA; excitation = 544 nm, emission = 590 nm). Bacterial viability was calculated as a percentage of resazurin turnover in the absence of compound. MICs of the clinical isolates were determined by means of the micro-broth dilution method. Dilutions of clinical isolates (about 106 CFU/mL) were streaked onto 7H11 solid medium containing a range of drug concentrations. Plates were incubated at 37 ◦C for about 5 days for MABSC or 7 days for MAC. The growth was visually evaluated: the lowest drug dilution at which visible growth failed to occur was taken as the MIC value. Results were expressed as the average of at least three independent replicates.

4.4. Analysis of the Lipid Composition of Mab Cells The analysis of lipid composition was performed by adapting the procedure previously reported for M. tuberculosis [16]. Briefly, Mab cultures were grown in Middlebrook 7H9 under standing conditions at 37 ◦C. After 24 h, mefloquine and NSC135792 were added to the cultures at a concentration of 25-fold MIC for 48 h. BM212, a known inhibitor of MmpL3, was used as a positive control. Then, cells were harvested by centrifugation and pellets were used for analysis. Cells were delipidated by extractions with 6 mL of chloroform/methanol (1:2) for 2 h, stirring at 56 ◦C, followed by two extractions with 6 mL of chloroform/methanol (2:1). Extracted lipids were subjected to biphasic washes with chloroform/methanol/water (4:2:1), dried under vacuum. Residues were resuspended in chloroform/methanol (2:1), in a volume of 100 µL for 30 mL of culture with an OD600 of 0.4. Samples (10 µL) were loaded on a TLC plate and developed in chloroform/methanol/water (20:4:0.5). Lipids were visualized with CuSO4 (10% in 8% phosphoric acid) and heating. Pellets from the same cultures were used to analyze the cell wall-bound mycolic acids. To this purpose, delipidated cell pellets were subjected to alkali saponification with 2 mL of 15% tetrabutylammonium hydroxide solution (TBAH) at 100 ◦C overnight. After cooling, 3 mL of dichloromethane, 2 mL of water, and 300 µL of iodomethane were added and stirred at rt, for 4 h. The mixtures were then centrifuged, aqueous phases were removed, and the organic phases were washed with water and dried under vacuum. For the extraction of mycolic acid methyl esters (MAMEs), samples were resuspended in 3 mL of diethyl-ether, sonicated, and after centrifugation, dried under vacuum. Extracts were then resuspended in chloroform/methanol (2:1), according to the OD600 of the culture as above, and 20 µL was loaded on TLC, developed in n-hexane/ethyl acetate (95:5, 3-fold), and visualized with CuSO4.

4.5. Checkerboard Assay To determine pairwise drug interactions, we used CBA assays. Briefly, serial dilutions of drug partners were performed at different combination ratios in a 96-well black plate (Fluoronunc, Thermo Fisher, Waltham, MA, USA); then, 100 µL of a bacterial inoculum, approximately of 106 CFU/mL, was added. Growth controls containing no compound and Int. J. Mol. Sci. 2021, 22, 8533 14 of 16

sterile controls without inoculum were also included. After 24 h of incubation at 37 ◦C, 10 µL of resazurin (0.025% w/v) was added to each well. The effect of the combination was measured using growth inhibition as the endpoint readout after a further 18–24 h of incubation, by means of a FluoroskanTM Microplate Fluorometer (Thermo Fisher Scientific, Waltham, MA, USA; excitation = 544 nm, emission = 590 nm). Then, the FICI value was calculated. The MIC of drug A in the presence of drug B divided by the MIC of drug A alone (FICA = [MICA(B)/MICA]) is defined as the FIC of drug A (FICA); and vice versa (FICB = [MICB(A)/MICB]). The sum of these values gives the final parameter FICI. “Synergy” is defined as a ≥4-fold reduction in the MICs of both compounds in combination compared to their MICs alone (FICI ≤ 0.5); “no interaction” when the MIC of one of the compounds 1 remained in the range of 2 × to 4× MIC (FICI > 0.5–4); and “antagonism” when the MIC of both compounds is, at least, 4-fold higher than that compared to the activity of the compounds alone (FICI > 4.0) [34].

Supplementary Materials: Supplementary Materials are available online at https://www.mdpi. com/article/10.3390/ijms22168533/s1. Author Contributions: Conceptualization, M.R.P. and F.M.; methodology, F.M., L.R.C., G.D., D.R., E.P., B.G., E.V.F. and L.F.; software, F.M.; validation, F.M., L.R.C., G.D. and M.R.P.; formal analysis, L.R.C., G.D., D.R., E.P. and B.G.; investigation, F.M., L.R.C., G.D., D.R., E.P. and B.G.; resources, M.R.P. and F.M.; data curation, F.M., L.R.C., G.D. and M.R.P.; writing—original draft preparation, F.M., L.R.C. and G.D.; writing—review and editing, F.M. and M.R.P.; visualization, F.M. and M.R.P.; supervision, F.M. and M.R.P.; project administration, M.R.P.; funding acquisition, M.R.P. and F.M. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by the Italian Cystic Fibrosis Foundation, grant number FFC#14/ 2020 (adopted by Smartform/Tattooform, Delegazione FFC di Belluno con i rocciatori di Fonzaso) (Giulia Degiacomi, Laurent Roberto Chiarelli, Deborah Recchia, Fabrizio Manetti and Maria Rosalia Pasca). This research was also funded by the Italian Ministry of Education, University and Research (MIUR): Dipartimenti di Eccellenza Program (2018–2022)—Dept. of Biology and Biotechnology “L. Spallanzani”, University of Pavia (Giulia Degiacomi, Laurent Roberto Chiarelli, Deborah Recchia and Maria Rosalia Pasca), and Dept. of Biotechnology Chemistry and Pharmacy, University of Siena (Fabrizio Manetti, Elena Petricci, and Beatrice Gianibbi). Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Acknowledgments: This work is dedicated to Gianni Mastella, who was the Scientific Director of the Italian Cystic Fibrosis Foundation; he devoted his life to the fight of cystic fibrosis (CF) and to the care of CF-affected individuals. We thank Stewart Cole (Pasteur Institute, Paris, France) and Claudia Sala (Fondazione Toscana Life Sciences, Siena, Italy) for providing NTM strains and M. abscessus MDR clinical isolate n. 1. We also thank Jan Madacki for the fruitful discussion about mycolates extraction. The Developmental Therapeutics Program, Division of Cancer Treatment and Diagnosis, National Cancer Institute, National Institutes of Health, is acknowledged for providing NSC compounds. Conflicts of Interest: The authors declare no conflict of interest.

References 1. Martiniano, S.L.; Nick, J.A.; Daley, C.L. Nontuberculous Mycobacterial Infections in Cystic Fibrosis. Thorac. Surg. Clin. 2019, 29, 95–108. [CrossRef][PubMed] 2. Salsgiver, E.L.; Fink, A.K.; Knapp, E.A.; LiPuma, J.J.; Olivier, K.N.; Marshall, B.C.; Saiman, L. Changing Epidemiology of the Respiratory Bacteriology of Patients with Cystic Fibrosis. Chest 2016, 149, 390–400. [CrossRef][PubMed] 3. Wu, M.L.; Aziz, D.B.; Dartois, V.; Dick, T. NTM drug discovery: Status, gaps and the way forward. Drug. Discov. Today 2018, 23, 1502–1519. [CrossRef] 4. Degiacomi, G.; Sammartino, J.C.; Chiarelli, L.R.; Riabova, O.; Makarov, V.; Pasca, M.R. Mycobacterium abscessus, an emerging and worrisome pathogen among cystic fibrosis patients. Int. J. Mol. Sci. 2019, 20, 5868. [CrossRef] 5. Daley, C.L.; Iaccarino, J.M.; Lange, C.; Cambau, E.; Wallace, R.J.; Andrejak, C.; Böttger, E.C.; Brozek, J.; Griffith, D.E.; Guglielmetti, L.; et al. Treatment of Nontuberculous Mycobacterial Pulmonary Disease: An Official ATS/ERS/ESCMID/IDSA Clinical Practice Guideline. Clin. Infect. Dis. 2020, 71, 905–913. [CrossRef][PubMed] Int. J. Mol. Sci. 2021, 22, 8533 15 of 16

6. Floto, R.A.; Olivier, K.N.; Saiman, L.; Daley, C.L.; Herrmann, J.L.; Nick, J.A.; Noone, P.G.; Bilton, D.; Corris, P.; Gibson, R.L.; et al. US Cystic Fibrosis Foundation and European Cystic Fibrosis Society consensus recommendations for the management of non-tuberculous mycobacteria in individuals with cystic fibrosis. Thorax 2016, 71, 88–90. [CrossRef] 7. Luthra, S.; Rominski, A.; Sander, P. The Role of Antibiotic-Target-Modifying and Antibiotic-Modifying Enzymes in Mycobacterium abscessus Drug Resistance. Front. Microbiol. 2018, 9, 2179. [CrossRef] 8. Egorova, A.; Jackson, M.; Gavrilyuk, V.; Makarov, V. Pipeline of anti-Mycobacterium abscessus small molecules: Repurposable drugs and promising novel chemical entities. Med. Res. Rev. 2021.[CrossRef] 9. Johansen, M.D.; Herrmann, J.L.; Kremer, L. Non-tuberculous mycobacteria and the rise of Mycobacterium abscessus. Nat. Rev. Microbiol. 2020, 18, 392–407. [CrossRef][PubMed] 10. Dupont, C.; Viljoen, A.; Dubar, F.; Blaise, M.; Bernut, A.; Pawlik, A.; Bouchier, C.; Brosch, R.; Guérardel, Y.; Lelièvre, J.; et al. A new piperidinol derivative targeting mycolic acid transport in Mycobacterium Abscessus. Mol. Microbiol. 2016, 101, 515–529. [CrossRef][PubMed] 11. Kozikowski, A.P.; Onajole, O.K.; Stec, J.; Dupont, C.; Viljoen, A.; Richard, M.; Chaira, T.; Lun, S.; Bishai, W.; Raj, V.S.; et al. Targeting mycolic acid transport by indole-2-carboxamides for the treatment of Mycobacterium abscessus infections. J. Med. Chem. 2017, 60, 5876–5888. [CrossRef][PubMed] 12. Franz, N.D.; Belardinelli, J.M.; Kaminski, M.A.; Dunn, L.C.; Calado Nogueira de Moura, V.; Blaha, M.A.; Truong, D.D.; Li, W.; Jackson, M.; North, E.J. Design, synthesis and evaluation of indole-2-carboxamides with pan anti- mycobacterial activity. Bioorg. Med. Chem. 2017, 25, 3746–3755. [CrossRef][PubMed] 13. Pandya, A.N.; Prathipati, P.K.; Hegde, P.; Li, W.; Graham, K.F.; Mandal, S.; Drescher, K.M.; Destache, C.J.; Ordway, D.; Jackson, M.; et al. Indole-2-carboxamides are active against Mycobacterium abscessus in a mouse model of acute infection. Antimicrob. Agents Chemother. 2019, 63, e02245–e02318. [CrossRef] 14. Li, W.; Yazidi, A.; Pandya, A.N.; Hegde, P.; Tong, W.; Calado Nogueira de Moura, V.; North, E.J.; Sygusch, J.; Jackson, M. MmpL3 as a target for the treatment of drug-resistant nontuberculous mycobacterial infections. Front. Microbiol. 2018, 9, 1547. [CrossRef] 15. Li, W.; Obregón-Henao, A.; Wallach, J.B.; North, E.J.; Lee, R.E.; Gonzalez-Juarrero, M.; Schnappinger, D.; Jackson, M. Therapeutic potential of the Mycobacterium tuberculosis mycolic acid transporter, MmpL3. Antimicrob. Agents Chemother. 2016, 60, 5198–5207. [CrossRef][PubMed] 16. Degiacomi, G.; Benjak, A.; Madacki, J.; Boldrin, F.; Provvedi, R.; Palù, G.; Kordulakova, J.; Cole, S.T.; Manganelli, R. Essentiality of mmpL3 and impact of its silencing on Mycobacterium tuberculosis gene expression. Sci. Rep. 2017, 7, 43495. [CrossRef][PubMed] 17. Zhang, B.; Li, J.; Yang, X.; Wu, L.; Zhang, J.; Yang, Y.; Zhao, Y.; Zhang, L.; Yang, X.; Yang, X.; et al. Crystal structures of membrane transporter MmpL3, an anti-TB drug target. Cell 2019, 176, 636–648. [CrossRef] 18. Dixon, S.L.; Smondyrev, A.M.; Knoll, E.H.; Rao, S.N.; Shaw, D.E.; Friesner, R.A. PHASE: A new engine for pharmacophore perception, 3D QSAR model development, and 3d database screening. 1. Methodology and preliminary results. J. Comput. Aided Mol. Des. 2006, 20, 647–671. [CrossRef] 19. Wishart, D.S.; Knox, C.; Guo, A.C.; Shrivastava, S.; Hassanali, M.; Stothard, P.; Chang, Z.; Woolsey, J. DrugBank: A comprehensive resource for in silico drug discovery and exploration. Nucleic Acids Res. 2006, 34, D668–D672. [CrossRef] 20. Developmental Therapeutics Program. Available online: https://dtp.cancer.gov (accessed on 31 January 2021). 21. Xu, Z.; Meshcheryakov, V.A.; Poce, G.; Chng, S.S. MmpL3 is the flippase for mycolic acids in mycobacteria. Proc. Natl. Acad. Sci. USA 2017, 114, 7993–7998. [CrossRef] 22. La Rosa, V.; Poce, G.; Canseco, J.O.; Buroni, S.; Pasca, M.R.; Biava, M.; Raju, R.M.; Porretta, G.C.; Alfonso, S.; Battilocchio, C.; et al. MmpL3 is the cellular target of the antitubercular pyrrole derivative BM212. Antimicrob. Agents Chemother. 2012, 56, 324–331. [CrossRef] 23. Barry, C.E., III; Lee, R.E.; Mdluli, K.; Sampson, A.E.; Schroeder, B.G.; Slayden, R.A.; Yuan, Y. Mycolic acids: Structure, biosynthesis and physiological functions. Prog. Lipid Res. 1998, 37, 143–179. [CrossRef] 24. Raynaud, C.; Daher, W.; Johansen, M.D.; Roquet-Banères, F.; Blaise, M.; Onajole, O.K.; Kozikowski, A.P.; Herrmann, J.L.; Dziadek, J.; Gobis, K.; et al. Active benzimidazole derivatives targeting the MmpL3 transporter in Mycobacterium abscessus. ACS Infect. Dis. 2020, 6, 324–337. [CrossRef][PubMed] 25. Kwak, N.; Dalcolmo, M.P.; Daley, C.L.; Eather, G.; Gayoso, R.; Hasegawa, N.; Jhun, B.W.; Koh, W.J.; Namkoong, H.; Park, J.; et al. Mycobacterium abscessus pulmonary disease: Individual patient data meta-analysis. Eur. Respir. J. 2019, 54, 180199. [CrossRef] 26. Degiacomi, G.; Belardinelli, J.M.; Pasca, M.R.; De Rossi, E.; Riccardi, G.; Chiarelli, L.R. Promiscuous Targets for Antitubercular Drug Discovery: The Paradigm of DprE1 and MmpL3. Appl. Sci. 2020, 10, 623. [CrossRef] 27. Sethiya, J.P.; Sowards, M.A.; Jackson, M.; North, E.J. MmpL3 inhibition: A new approach to treat nontuberculous mycobacterial infections. Int. J. Mol. Sci. 2020, 21, 6202. [CrossRef][PubMed] 28. Graham, J.; Wong, C.E.; Day, J.; McFaddin, E.; Ochsner, U.; Hoang, T.; Young, C.L.; Ribble, W.; DeGroote, M.A.; Jarvis, T.; et al. Discovery of benzothiazole amides as potent antimycobacterial agents. Bioorg. Med. Chem. Lett. 2018, 28, 3177–3181. [CrossRef] 29. Ashley, E.A.; Phyo, A.P. Drugs in Development for Malaria. Drugs 2018, 78, 861–879. [CrossRef] 30. Schrodinger Release. Schrödinger Release 2019-1; Maestro, Schrödinger, LLC: New York, NY, USA, 2019. 31. Shelley, J.C.; Cholleti, A.; Frye, L.; Greenwood, J.R.; Timlin, M.R.; Uchimaya, M. Epik: A software program for pKa prediction and protonation state generation for drug-like molecules. J. Comp. Aided Mol. Des. 2007, 21, 669–681. [CrossRef][PubMed] Int. J. Mol. Sci. 2021, 22, 8533 16 of 16

32. Friesner, R.A.; Banks, J.L.; Murphy, R.B.; Halgren, T.A.; Klicic, J.J.; Mainz, D.T.; Repasky, M.P.; Knoll, E.H.; Shaw, D.E.; Shelley, M.; et al. Glide: A new approach for rapid, accurate docking and scoring. 1. Method and assessment of docking accuracy. J. Med. Chem. 2004, 47, 1739–1749. [CrossRef] 33. Palomino, J.C.; Martin, A.; Camacho, M.; Guerra, H.; Swings, J.; Portaels, F. Resazurin microtiter assay plate: Simple and inexpensive method for detection of drug resistance in Mycobacterium tuberculosis. Antimicrob. Agents Chemother. 2002, 46, 2720. [CrossRef][PubMed] 34. Gómara, M.; Ramón-García, S. The FICI paradigm: Correcting flaws in antimicrobial in vitro synergy screens at their inception. Biochem. Pharmacol. 2019, 163, 299–307. [CrossRef][PubMed]